Competitor analysis on Ozon: reading what the other sellers are doing

This article assumes the niche is already chosen. It is about reading the sellers next to you once you are trading there: which lines they check on a competitor's card, how prices get monitored, where the price index pushes, how dumping gets diagnosed, and which tools get named for the job. Every concrete claim below comes from the transcript of one of 18 Russian-language seller videos. None of them is an official Ozon document; two are sponsored tool reviews that hand out a discount code. The figures belong to the period each video was recorded — the transcripts were collected in July-August 2026, the videos themselves span 2024-2026. Ozon's subscription packages and panel screens change, so treat the thresholds as orientation, not rules.

What sellers actually read on a competitor's card

Age first. In the selection criteria Nikita describes, the ideal competitor has no more than 500 reviews and has not been selling on Ozon for longer than half a year; anything older is hard to repeat, because you end up fighting cards that have traded for five years and carry 5,000 to 10,000 reviews. The competitor revenue they look for is 300,000 to 500,000 roubles on that product. Артём uses a faster version of the same test: few reviews and a card that has only been in stock for seven days means the listing really is new. Рома Райд builds the filter from the other end — more than 30 reviews in the last 30 days proves the item is selling right now, while fewer than 100 reviews in total means it has only just started.

Then the price curve. In his example a floor door stopper looks perfect — no reviews a month ago, reviews now — but the price is heading down, and that is a red flag. Another video describes the shape to look for: a couple of competitors, none of them dumping, no downward price graph, and at least one of them a newcomer, because it is far easier to inherit the experience of someone who started a month ago with about 30 reviews than of someone sitting on 3,000.

The column Ozon recently added settles a lot of arguments: the competitor's share of advertising cost on that product. Nikita uses it directly — you can now see in advance what a rival spends on ads at the price he sells for, and the question becomes blunt: if the competitor is at 25 per cent, why would yours be different? The same conversation mentions cases where that share reaches 40 per cent. The remaining lines of the card: below a 4.8 rating a listing sells badly, 4.7 is their floor, 4.9 is good. Reviews double as a content brief — in the orthopaedic pillow example the complaints are high price, clumping filler, uncomfortable sleep and poor quality, and those four headings go straight into your own card as answers.

One less obvious line: the sales-by-warehouse breakdown on a competitor's card. The course builds its whole supply plan on it — open the top five to ten cards in the category, copy each one's warehouse table, map warehouses onto clusters, and average the resulting regional percentages with your own non-local orders and with Маяк's supply calculation.

Before any of that there is one question: does your product compete with everyone? Рома Райд draws the line sharply. A 7.5 cm drain filter competes only with filters of the same diameter — a buyer looking for 7.1 will not buy 11 cm — so the real competition is the single seller who stocks that size, and that seller shifted a batch of 100 in a week. Children's smart watches have no one-to-one rival, but the buyer types a generic query and there everything competes with everything: even if the unit economics work for both products, the watch fight is fought with advertising.

Price monitoring and the pressure of the price index

The pricing videos share one hard sentence: on a marketplace, price decides whose product the buyer picks, and the winner is usually the one who is 20 roubles cheaper. Рома Райд says exactly one product in his own store sells 20 per cent above competitors; every other attempt to win on design and reviews simply never worked. His method: take the competitor whose listing you found the product on and run the unit economics at his price — does he even have net profit? The price on screen is not reliable either; in his example the item shows 424 roubles, but since it is unclear whether a discount is running he calculates with 324 and notes that the calculator does not show the real price.

The true competitor list is built by hand: type the query word into an analytics service, filter products by name, tick only the ones genuinely selling the same thing, export, and average the price. Then split the field into price segments and work out the unit economics of each segment together with how many sales competitors make inside it. The rule that falls out of this is the most practical line in the whole topic: do not sit far above the bottom segment. In the example some sell at 268 roubles and some at 452 — if you plan 452 while a dozen sellers are at 260, think ten times, and calculate your unit economics at the minimum segment too, because one day you may be forced down to it. Another video takes the shortcut: average 20 competitors and price at, or slightly above, that market average.

On frequency the videos diverge. One recommends recalculating once or twice a month, because the exchange rate, purchase cost, commissions and cost structure all keep moving. Рома Райд says they control prices practically every day. The course instructor describes opening the index every morning and cutting prices where needed. One detail shows how fast the field reacts: when they launch a product and appear in the new-arrivals ranking, more than 50 sellers follow them into that product.

The real pressure sits in the price index (индекс цен). One video describes the mechanism in three stages: first the lowest competitor price inside Ozon, then the lowest price of your own product outside Ozon, then the lowest price of a competitor's equivalent on other platforms. In the second and third stages Ozon looks at the price the buyer sees — including a discount that the other platform funds itself, so a discount you never gave drags you down. The numbers: the index runs from 0.16 to 1.99; up to 1.1 is green, 1.1 to 1.5 is amber, anything higher is red. The formula is simple — if your price is lower than the competitor's, divide yours by his; if it is higher, the result is 2 minus his price divided by yours. Mis-indexing examples are given too: Ozon finds a cheap Chinese equivalent on another platform, or compares a single unit against your two-piece set, since one toothbrush is obviously cheaper than two, and support often does not fix it.

On what the index actually does the videos disagree. We are not resolving the contradiction, only reporting it:

The only pair of figures that recurs in two independent videos is this one: on a favourable index, Premium +10 per cent and Premium Plus +12.5 per cent. The badge condition is stated precisely — if an index is assigned on more than half of your products, more than 80 per cent of those are green and fewer than 10 per cent are red, Ozon may put the badge on your store card. A bad index, by contrast, is said not to affect the discount Ozon funds from its own pocket, which depends on promotion participation, brand reputation and other factors.

Sellers answer the index on three levels. The innocent one sits in the panel: under products and prices you can add links to competitor listings that improve your own index. The second is a cross-platform game, which the course lays out as three strategies: the ladder (lowest price on one marketplace, middle on Ozon, highest on another, with Ozon's index tied to the expensive one), cross-platform (the same goods under two brands, each looking favourable where it matters), and nothing to do with me (change the visual, one characteristic and the name, then write to support that the items differ and cannot be compared). The third level is openly against the rules: one video suggests listing the same item elsewhere 10 per cent higher, handing Ozon that link, and adding a few fake self-purchases to give the listing weight — in the presenter's own words, bending the rules a little. We report it, we do not recommend it: self-purchases breach marketplace rules.

The same video notes the political side. In February, entrepreneurs who lost real money to faulty index calculations complained to Russia's competition authority, but Ozon sits in the register of IT companies exempted from antitrust supervision until the end of that year, so he expects nothing. His constructive advice is not to fight the index at all: build brand reputation and drive returns down, because a genuinely loyal buyer will not go to a stranger just for a better price.

Is the niche monopolised? The numbers sellers use

The first two ratios are always the same: the share of products with sales and the share of sellers with sales. Артём checks both the moment he opens a niche; 54 per cent counts as low and the band he wants is 70 to 75 per cent and above. In his example a cable-lug category passes at 80 and 87 per cent. Then comes the monopolisation analysis — three pie charts showing how revenue is split across products, brands and sellers. A healthy reading: products outside the top take 78 to 84 per cent of revenue and sales spread fairly evenly across sellers. A broken one: a single product at 16.4 per cent, a single brand at 23 per cent. In the barbecue category one brand holds 13 per cent, another 10, and the remaining brands 31.7; on the seller side it is 13, 10.8 and 33.5 per cent for the rest.

Another video does the same job with MoneyPlace. The service scores a niche out of 15 for prospects; women's dresses score 4 out of 15. The niche card carries market volume, revenue sellers are missing, the top-1 share, the top-10 share (24 per cent in the example), the proportion of products with sales, average sales speed, average revenue per seller, average rating and review count. His own criteria come down to three lines: how many products the query returns, query frequency, and how many competitors there are. Concretely — a micro-earphone with 54,000 query frequency against only 962 products; a newborn cocoon with 89,000 queries against 1,000 products and 54 per cent of sales organic, where in most niches traffic is almost entirely paid.

Дмитрий Хван reduces the same decision to four questions using a free browser extension. One: is the revenue on the first page of the highest-frequency query above 40 million roubles (20 to 60 million is fine with 200,000 to 500,000 roubles of capital). Two: is one brand, or a handful, taking the money on that first page? Laundry gel shows 196 million roubles on the first page split between four brands — a monopoly, red light; nappies look the same. Three: can you make the product distinctive? Four: total product count above 25,000 and he stays out, while 10,000 is still normal.

A podcast guest measures the same thing as density rather than turnover: you open Ozon analytics, see billions in monthly orders in clothing and footwear, and it is nonsense. Bags run about 1.2 billion roubles a month in orders against roughly 40,000 sellers; divide one by the other and you can see why you should stay away. He adds that in bags the dumping is brutal and the copying is brutal, and advises avoiding niches with many sizes and colours because the frozen capital is enormous. Рома Райд looks at evenness instead: have a few competitors, but let turnover be split roughly equally — if one of them does 100,000 and the rest do 10,000, you will be joining the 10,000 group.

The one source that pushes back deserves recording too. Nikita is asked whether monopolists in a niche matter when choosing a product; his answer is that monopolists exist in every niche and you have to test. What follows is more interesting: the assumption that broke for them that season was that unbranded small appliances — kettles, blenders, waffle irons — would not sell against the famous names. When they tried it, a single card did 2 to 5 million roubles with good profitability. People buy even when they do not know the brand, he says, and after that they started launching everything, televisions included. In the same interview he says his advertising costs on Ozon are enormous and he is considering leaving the platform.

Dumping: how it is spotted and what sellers do

The definition comes from an SEO case study: competitors sell at 2,000, you enter the same product at 1,500 and trade at zero or a small loss, expecting to spin things up quickly. What it looks like on the shelf comes from Рома Райд's dish-sponge example. The unit economics clear 20 per cent margin, but the category leader has 5,000 reviews, a pack of ten sits at 334 roubles with the price falling, there are many identical competitors, four pieces go for 100 roubles and someone has dropped to 68 — either he wants rid of them or he is selling them as an economy line. New sellers with a similar item have been trying for a month; one who entered in early August sold almost nothing in three months, because he skipped competitor analysis. The conclusion is flat: competing with sellers like that is practically impossible.

The clearest answer on where dumping comes from arrives inside the China debate. In the screw-hook category the entire top 12 is local sellers and the Chinese listings barely sell, yet the listing is still burnt out. The cause is not the Chinese: there are more products and more sellers than there are real buyers willing to pay a given price, and that is where dumping starts. The practical test follows: if there are no Chinese sellers in the top 12, do not worry — his claim is that China does not affect roughly 95 per cent of categories, and that you only need to avoid products where Chinese sales clearly exceed local ones.

Automated dumping is a separate species. A guest selling through Ozon Global says there were over 100,000 Chinese seller accounts on Ozon in 2023 and clearly more now, and that about two thirds of the listings filling the platform are dropshipping systems. They clone cards through the API, recognise the image, buy the equivalent on Chinese marketplaces and ship it without a human touching it. Because they cannot use advertising tools, points for reviews or bonus mailings, the only lever they have is the minimum price. The metaphor fits: a normal seller is an angler going deep with one card, they are fishing with a net.

The cost of dumping yourself is described similarly in two videos. First: the profit you gave up cannot be reinvested into growth, packaging and quality, and the way back is expensive — returning from 1,500 to the 1,700-1,800 band takes a month to six weeks, and getting back to 2,000 with an acceptable volume takes two to three months. He does not recommend dumping and does not use it in his own store. The second is harsher: climbing out takes several months up to half a year, and some never manage it, because Ozon's algorithm has got used to selling that product at that price; when the seller raises it, the card collapses to position 200, 300 or even 500. The single accepted exception is a conscious strategy — starting at minimal profit to accumulate orders and reviews, then returning to normal prices over months. Switch dumping on blindly and expect sales, and it simply does not work.

The alternatives are grouped into three in the course: hold the lowest price inside your own price segment, since buyers filter by segment; sell at the average market price; or tune the advertising bid so you land among competitors whose prices are higher, where you look cheap by comparison. With a caveat: on another marketplace the bid-to-position relationship can be calculated, on Ozon it cannot, only tested. And the blunt reality everyone accepts: if the buyer can get the same item even 10 per cent cheaper, he will — nothing personal, just business. Finally, the diagnostic value of watching rivals: when orders in Nikita's team fell from 15 to 2, the first question was whether the competitor still had 15. Because they track all competitors, they can separate a platform-wide change from their own mistake.

The tools they name

The single most valuable find in these videos is Ozon's own report. Under seller analytics, products in search, sits the beta report «Что влияет на место в поиске»: you pick your own article, the buyer's region and a search query, and Ozon returns your position, how favourable your price is considered and how many boosting points that earns, your popularity on the query, and the positions and scores of competitors. The course instructor calls this report his competitor analysis outright and stresses that it is free — no Premium subscription required. The other three reports he uses regularly are the sales funnel (impressions, unique visitors who viewed the card, cart conversion, purchased out of ordered), position in search (the correlation between position and orders, tested by moving the ad bid) and the effect of promotions on sales.

The shared weakness of external services is stated openly: their data does not reconcile with Ozon's own numbers, because Ozon keeps the data inside. The course instructor offers one cross-check: Ozon showed a query popularity of 30,770 for one smartphone query while MPStats reported 32,000 a month for its plural form — close enough, in his words. There is also a rule of thumb for subscriptions: Premium should never exceed 1 per cent of your unit-economics costs. That maps onto monthly turnover — Light up to 500,000 roubles, ordinary Premium from a million, and Plus only really justified from 5 to 6 million. The same video warns that buying position is pointless if it zeroes your margin, and notes that the top 10 positions take about 90 per cent of sales and the top 5 about 80 per cent.

Short summary

• On a competitor's card, age is read first: under 500 reviews, a listing younger than six months, a price curve that is not falling.
• Ozon's new column shows a competitor's advertising cost share; if he is at 25 per cent, do not assume you will be different.
• Unit economics are calculated first at the competitor's price, then against the lowest price segment.
• The price index is computed from the lowest price inside and outside Ozon; green up to 1.1, red above 1.5.
• The videos contradict each other on whether the index affects ranking; the only figures recurring in two independent sources are Premium +10 and Plus +12.5 per cent.
• Monopolisation is measured with three numbers: products and sellers with sales, who owns the first page, revenue divided by seller count.
• Dumping usually starts because there are more sellers than buyers at a given price; climbing out takes two to three months, sometimes six.
• Ozon's free beta report shows competitors' boosting scores — external services cannot see that.